Structural and Biochemical Characterization of Aal, a Quorum

www.nature.com/scientificreports

OPEN Structural and Biochemical Characterization of AaL, a Quorum Quenching Lactonase with Unusual Received: 1 March 2018 Accepted: 29 June 2018 Kinetic Properties Published: xx xx xxxx Celine Bergonzi1, Michael Schwab1, Tanushree Naik1, David Daudé 2, Eric Chabrière3 & Mikael Elias1

Quorum quenching lactonases are enzymes that are capable of disrupting bacterial signaling based on acyl homoserine lactones (AHL) via their enzymatic degradation. In particular, lactonases have therefore been demonstrated to inhibit bacterial behaviors that depend on these chemicals, such as the formation of bioflms or the expression of virulence factors. Here we characterized biochemically and structurally a novel representative from the metallo-β-lactamase superfamily, named AaL that was isolated from the thermoacidophilic bacterium Alicyclobacillus acidoterrestris. AaL is a potent quorum quenching enzyme as demonstrated by its ability to inhibit the bioflm formation of Acinetobacter baumannii. Kinetic studies demonstrate that AaL is both a profcient and a broad spectrum enzyme, 5 −1 −1 being capable of hydrolyzing a wide range of lactones with high rates (kcat/KM > 10 M .s ). Additionally, AaL exhibits unusually low KM values, ranging from 10 to 80 µM. Analysis of AaL structures bound to phosphate, glycerol, and C6-AHL reveals a unique hydrophobic patch (W26, F87 and I237), involved in substrate binding, possibly accounting for the enzyme’s high specifcity. Identifying the specifcity determinants will aid the development of highly specifc quorum quenching enzymes as potential therapeutics.

Bacterial quorum sensing (QS) is one of the most prominent communication system displayed by bacteria. Tis system depends on the production and detection of signaling molecules referred to as “autoinducers” to coordinate gene expression in response to cell density. A common class of autoinducers are N-acyl-L-homoserine lactones (AHLs)1. AHL-based bacterial quorum sensing was shown to regulate various behaviors in numerous microbes, mainly gram-negatives, including virulence factors production and bioflm formation2. Some enzymes, named Quorum Quenching (QQ) enzymes, are naturally capable of interfering with AHL-based QS, via the enzymatic modifcation or degradation of the signaling molecules3. Consequently, these enzymes were previously reported as bioflm and virulence inhibitors4 during in vitro and in vivo experiments5–8. Tey comprise promising candidates as potential protein therapeutics9, to prevent infection in livestocks10 and to prevent biofouling11. AHL-degrading enzymes can be found in organisms beyond the bacterial world, e.g. archae, plants, fungi and mammals12,13. Some of the most widely used and studied QQ enzymes are lactonases. Lactonases were identifed and characterized from three main protein superfamilies14. Te paraoxonases (PONs) are lactonases that were primarily identifed in mammals13,15,16. PONs adopt a six-bladed β-propeller fold and a central tunnel with two calcium cations17,18, one being structural, the second being involved in catalysis19. PONs were shown to hydrolyze δ-lactones, γ-lactones and AHLs20. Another class of lactonases is the Phosphotriesterases-like lactonase (PLL). Several PLLs were identifed in several extremophiles21–23 and accordingly exhibit remarkable thermal stability (up to 128 °C for VmoLac22) and high rates for lactone hydrolysis22. PLLs were divided into two subclasses, A and B, where PLL-As can hydrolyze δ-lactones, γ-lactones and AHLs and PLL-Bs prefer δ-lactones and γ-lactones22. Notably, certain γ-lactones can be used as QS molecules in Streptomyces and Rhodococcus24,25.

1Biochemistry, Molecular Biology & Biophysics Dpt and BioTechnology Institute, University of Minnesota, Saint Paul, Minnesota, 55108, USA. 2Gene&GreenTK, 19–21 Boulevard Jean Moulin, 13005, Marseille, France. 3Aix Marseille Univ, IRD, APHM, MEPHI, IHU-Méditerranée Infection, Marseille, France. Correspondence and requests for materials should be addressed to M.E. (email: [email protected])

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 1 www.nature.com/scientificreports/

Kcat KM Kcat/KM Enzyme Organism Substrate (s−1) (µM) (M−1 s−1) Reference AaL Alicyclobacillus acidoterrestris C4-AHL 12.6 9 1.3 × 106 Tis work* AiiA Bacillus thuringiensis C6-AHL 91 5600 1.6 × 104 26 AiiB Agrobacterium tumefaciens C6-AHL 24.8 1600 1.6 × 104 27 MBLs Chryseobacterium sp. strain AidC C7-AHL 80 47 1.7 × 106 28 StRB126 MomL Muricauda olearia 3-oxo-C10-AHL 224 440 5.1 × 105 29 PON2 Mammals 3-oxo-C12-AHL 13 50 2.7 × 105 20 PONs PON-OCCAL Mammals 3-oxo-C12-AHL 44.5 180 2.4 × 105 20 SsoPox Sulfolobus solfataricus 3-oxo-C10-AHL 4.5 143 3.6 × 104 36 VmoLac Vulcanisaeta moutnovskia 3-oxo-C10-AHL 0.5 231 2.1 × 103 22 PLLs SacPox Sulfolobus acidocaldarius C8-AHL 0.9 178 5.3 × 103 77 SisLac Sulfolobus islandicus 3-oxo-C8-AHL 4.1 42 9.7 × 104 23 Acylase PvdQ Pseudomonas aeruginosa C12-AHL 2.5 11 2.2 × 105 78

Table 1. Kinetics parameters for several quorum quenching enzyme representatives. *Te kinetics were performed in activity bufer (2.5 mM Bicine pH 8.3, 150 mM NaCl, 0.2 mM cresol purple, 0.5% DMSO and with 0.2 mM CoCl2).

Te metallo-β-lactamase like (MLLs) or AiiA-like represents another class of lactonases. Te most characterized enzyme from this family is the autoinducer inactivator A (AiiA) from Bacillus thuringiensis26, but other representatives have been studied, such as AiiB27, AidC28, MomL29 or GcL30. Te MLLs enzymes possess the conserved motif “HXHXDH” that allows for the formation of the bi-metallic active site. Its crystal structure has been solved31 and its catalytic mechanism has been investigated26,32. Numerous experiments in vivo and in vitro showed the efciency of these enzymes to degrade the quorum sensing signals, through the inhibitions of virulence factor production or bioflm synthesis for examples3,33. MLLs, including AiiA, AidC, MomL, were reported to exhibit broad substrate range with respect to the acyl chain length of AHLs4, but their potential activity on δ-lactones or γ-lactones is unknown. AaL (WP_021296945.1) is a lactonase isolated from the acidophilic, moderately thermostable bacterium 34 Alicyclobacillus acidoterrestris. Consequently, AaL is moderately thermostable, with a Tm of 58.2 °C . It shares 27% sequence identity with AiiA and 43% sequence identity with AiiB. Previous kinetic studies on AaL revealed that it is a very profcient lactonase, with broad specifcity spectrum, and also that it exhibits unusually low KM 34 values as compared to other MLLs lactonases . Indeed, KM values for AaL were found to range between 10 to 34 83 µM with AHLs as substrates, while several other MLLs were reported to exhibit much higher KM values (~1 mM for AiiA26,32, 440 µM for MomL29) with the exception of AidC (46–72 μM35). In AidC, its high specifcity was attributed to atypical structural features in the vicinity of the active site28. Other classes of lactonases 20,22,36,37 (e.g. PONs and PLLs) were reported to possess relatively low KM values (approximately 50 to 500 μM ) (Table 1). Low KM values enzymes are of particular interest because their study may allow for a better understand- ing on the structural determinants for lactones binding, and may represent promising tools and targets for their optimization as therapeutic, anti-virulence and anti-bioflm enzymes. Terefore, we undertook the structural characterization of the purifed AaL. We show that AaL is capable of degrading δ-lactones and γ-lactones with high catalytic profciency (>104 M−1 s−1) and shows some weak, promiscuous phosphotriesterase activity on the insecticide-derivated paraoxon (Fig. 1). Te crystal structure of AaL reveals a unique hydrophobic patch which could relate to its high specifcity, and highlights a fexible active site loop which may be involved in substrate binding. Results and Discussion Sequence analysis. Te AaL protein sequence was aligned with sequences of MLLs with known structures using MUSCLE program38 (Fig. 2). AaL is closer to AiiB (43% sequence identity), whereas it shares only 21 and 27% sequence identity with AidC and AiiA, respectively. All four enzymes possess the characteristic HXHXDH motif, as well as the two other residues (H and D) involved in the metals cations coordination. Other residues lining the active site do not show signifcant conservation, with the exceptions of Y223 (Fig. 2). Indeed, Y223 is conserved in all the known MLLs, with the exception of AidC35 where it is substituted by a His. Interestingly, this residue is also conserved in PLLs13,21, and has been proposed to be implicated in the catalytic mechanism31,39.

AaL is a proficient AHLs and lipophilic lactone hydrolase. AaL shows a wide activity spectrum against AHLs, as it hydrolyzes with similar, high catalytic profciencies (up to 106 M−1 s−1) AHLs with short, medium or long acyl chains (Table 2). Tis behavior has been previously observed for MomL29, AiiA40 and GcL30. Additionally, AaL is also able to degrade the γ- and δ-lactones with high profciency. While this feature was not reported for MLLs, PLL-As, but not PLL-Bs, are also known to hydrolyze both AHLs and lipophilic lac- tones21,23,36,41,42. Interestingly, γ-butyrolactone derivatives are known autoinducers for Streptomyces sp., potentially 43 suggesting a biological function to this activity . Remarkably, the KM values for AaL are low, as compared to most MLLs, and ranges between 9–83 μM with the exception of γ-butyrolactone (Fig. S3). Additionally, we report that AaL exhibits some promiscuous, weak organophosphorus hydrolysis activity with −1 −1 a kcat/KM value of 22.9 M s using paraoxon as a substrate. Tis trait has not been reported for other MLLs, but

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 2 www.nature.com/scientificreports/

Figure 1. Structures of the tested chemicals. (A) Paraoxon-ethyl (B) Acyl-Homoserine Lactones, (C) 3-oxo- Acyl-Homoserine Lactones, (D) γ-lactones and (E) δ-lactones

has been repeatedly described for PLLs. Some PLLs members were in fact identifed by virtue of their phosphotriesterase activity21,44,45. It has been hypothesized that lactonases, owing to their ability to hydrolyze organophosphorus compounds, are the progenitors of optimized phosphotriesterases46.

AaL is a quorum quenching lactonase. Acinetobacter baumannii is a human pathogen that is known to produce and utilize acyl homoserine lactone, including C12-AHL47. Quorum quenching enzymes, and in particular lactonases, were recently demonstrated to inhibit its bioflm formation6,48, and even disrupt existing bioflms49. Here we performed a bioflm inhibition dose-response experiment (Fig. 3). Tis experiment shows that AaL can inhibit the formation of bioflm, up to >4-fold as referred to untreated cultures, and >2.2-fold compared to an inactive lactone control (inactive mutant SsoPox 5A8; carrying the mutations V27G/P67Q/L72C/Y97S/Y99A/ T177D/R223L/L226Q/L228M/W263H, obtained previously50). Interestingly, and as previously observed for the lactonase SsoPox51,52, AaL has no negative efects on growth, and exhibits similar growth levels as the controls performed with Bovine Serum Albumine (BSA) and 5A8. Control made with the inactive lactonase variant 5A8 suggests that the observed bioflm inhibition depends on catalytic activity. Tis observation contrasts with the use of the previously reported Quorum Sensing Inhibitor (QSI) 5-FU53 that inhibits both growth and bioflm formation.

Crystal structure of AaL. Structures of AaL were solved in C2 space group and in two diferent unit cells parameters (Table 3). Te monomer of AaL is roughly globular with overall dimension of 60 Å × 52 Å × 43 Å approximately, and shows a long protruding loop (Fig. 4A). Tis loop is involved in homodimerization (Fig. 4B). Te homodimer measures approximately 94 Å × 52 Å × 43 Å. As expected, AaL exhibits a αβ/βα sandwich fold, typical to the metallo-β-lactamase superfamily, and is similar to that of others MLLs such as AiiB or AiiA26,54. Te overall structure is similar to the structure of AiiB (2R2D) from Agrobacterium tumefaciens (43% sequence identity). Te root-mean-square deviation (r.m.s.d) for α-carbon atoms (over 273 atoms) is 0.95 Å. AaL shows highly signifcant diferences with AiiA with a r.m.s.d for α-carbon atoms (overs 181 atoms) at 1.22 Ǻ. One of the most important diference between the structure of AiiA and AaL resides in an external loop involved in the dimerization. Tis external loop is produce by an insertion of nine amino acids between the residues P33 and A41 (Fig. 2). Another noticeable diference between AiiA and AaL resides in the access to the active site. In fact, AaL active site appears much more accessible than AiiA’s active site, which is partially obstructed by two residues (E135 and F68 in AiiA) (Fig. 5A,C). AaL and AidC shares 21% of sequence identity, and despite exhibiting a similar topology, their structures show major diferences. Te most important diferences are localized in the active site area, with a diferent orientation of the active site crevice28, as well as a diferent homodimerization surface28 (Fig. 4G,H). Te fnding that AaL, isolated from a thermoacidophilic bacteria, and contrary to AiiA14, is organized as a homodimer, is consistent with previous work on thermophilic proteins, highlighting a trend for higher levels of

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 3 www.nature.com/scientificreports/

Figure 2. Sequence alignment of the MLLs representatives. Sequence alignment of AaL from Alicyclobacillus acidoterrestris (WP_021296945.1), AiiA from Bacillus thuringiensis (PMC1187999), AiiB from Agrobacterium tumefaciens (WP_010974862.1) and AidC from Chryseobacterium sp. StRB126 (PMC4681436). Te residues corresponding to the hydrophobic patch (W26, F87 and I237) identifed in AaL are highlighted in pink and the amino acids involved the metal coordination (including the characteristic HXHXDH motif) are designated by a red star. Te position highlighted in green (Y223 in AaL) corresponds to a tyrosine residue possibly involved in the catalytic mechanism. Te amino acids in red are involved in the dimerization of AaL.

−1 −1 −1 Substrates kcat (s ) KM (μM) kcat/KM(M s ) C4- AHL (l)* 12.6 ± 0.8 9.4 ± 2.5 (1.3 ± 0.4) × 106 C6- AHL (l)* 14.0 ± 0.4 82.7 ± 11.0 (1.7 ± 0.2) × 105 C10- AHL (l)* 5.2 ± 0.3 57.2 ± 17.4 (9.2 ± 2.8) × 104 3-Oxo-C12-AHL (l) * 4.9 ± 0.3 17.4 ± 4.8 (2.8 ± 0.8) × 105 γ-butyrolactone 23.5 ± 1.7 171.0 ± 36.1 (1.4 ± 0.3) × 105 γ-heptanolide 7.0 ± 0.3 36.9 ± 8.9 (1.9 ± 0.4) × 104 δ-valerolactone 8.4 ± 0.4 18.7 ± 0.4 (4.5 ± 0.9) × 105 δ-decalactone 8.4 ± 0.6 18.5 ± 0.4 (4.5 ± 1.0) × 105 Paraoxon-ethyl ND ND 22.9 ± 1.4

Table 2. Enzymatic characterization of the AaL enzyme. Related chemical structures of substrates are presented in Table S1. Kinetics were measured as quadruplicates, and standard deviation values are given for each parameter. Te kinetics were performed in activity bufer (2.5 mM Bicine pH 8.3, 150 mM NaCl, 0.2 mM cresol purple, 0.5% DMSO and with 0.2 mM CoCl2). ND corresponds to not determined values because of too low catalytic rate. *Data from34.

oligomerization in these proteins55. Te dimer is characterized by a strong interaction of the protruding loop (P33 to A41) from both monomers (Fig. 4B). Te dimer interface involves 32 residues in each monomer. Te interface is mostly hydrophobic, and involves only 14 hydrogen bonds. Te contacting area is 1125.7 Å2, a similar value to other dimeric MLL structures such as AiiB and AidC, 1089.1 Å2 and 1015.4 Å2, respectively.

Active site of AaL. AaL possesses a bi-metallic active site center. Te bound metal cations nature in the MLLs was previously attributed to two zinc cations in AiiA26 as well as AiiB54. Metal cations are bound via fve

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 4 www.nature.com/scientificreports/

Figure 3. Efect of AaL on bioflm formation by Acinetobacter baumannii (A) and cell density (B) Bioflm was quantifed using crystal violet staining and amount of bioflm is expressed as absorbance (550 nm). Te cell density is a measure of cell growth in suspension measured as absorbance at 600 nm. Te inactive mutant of the lactonase SsoPox (5A8) containing 10 mutations (V27G/P67Q/L72C/Y97S/Y99A/T177D/R223L/L226Q/ L228M/W263H) and Bovine Serum Albumin (BSA) were used as negative controls while quorum sensing inhibitor (QSI) 5-fuoro-uracil has been used as positive control. Data shown are average of three independent experiments with 4 technical replicates each (n = 12), error bars represent the standard deviation. Student t-test results are shown (A) between data for bioflm formed with diferent doses of AaL and data for bioflm formed in presence of the inactive lactonase 5A8.

histidine residues (118, 120, 123, 198 and 266), and two aspartic acid residues (122 and 220). Anomalous X-ray scattering data collected at a higher energy than the Co-Kedge (7.7089 keV) shows that the active site may be occupied by cobalt cations, but not by other common metal cations identifed in similar enzymes such as zinc (Zn-K edge is 9.6586 KeV) or nickel (Ni-K edge is 8.3328)21,56 (Fig. S2). Tis result contrasts with known enzymes from the MLL family that were described to possess two zinc cations in their active site35,39,54 (Fig. S2), but is consistent with the fact that cobalt cations were added to the protein production steps. Terefore, cobalt was modelled in the active site of AaL. Te crystal structure of AaL reveals that its active site is hydrophobic. Besides tyrosine 223, a residue conserved in lactonases and potentially involved in the catalytic mechanism31,32,54,57, the active site cavity is decorated by the side chains of three methionine residues (20, 22 and 86), a tryptophan residue (26), a phenylalanine residue (87), an isoleucine residue (237), a leucine residue (121) and an alanine residue (157). Te hydrophobic character of the clef is reinforced by W200 and Y239 that might interact with the side chains of very long chain AHLs.

Comparison with AiiA. Both enzymes difer signifcantly in their active site accessibility. While AiiA has a relatively narrow active site entrance, AaL has a much more accessible active site center (Fig. 5A,C). Tis is due to diferent conformations of loops decorating the active sites of both enzymes, mainly caused by an insertion of seven amino acids (Y149 to E155) in AaL as compared to AiiA (Figs 2 and S4). Another diference relates to a loop surrounding the active site and containing I237: AaL exhibit a 2-residues insertion next the active site (P235 and G236) as compared to AiiA. Tis insertion afects the position of I237 which is lining the active site (Fig. 6A), and may alter the relative binding of lactones onto these two enzymes (Fig. S5). Lastly, whereas the key active site amino acids are well superposed (Fig. 6A), Y223 (in AaL) and the equivalent residue in AiiA (Y194) adopt a twisted orientation (39.9°). Given the possible implication of this residue for catalysis31,32,54,57, this may have implications on the role of this residue in both enzymes (Fig. 6A).

Comparison with AiiB. While AaL and AiiB share very similar structures, some diferences can be seen while comparing their respective active site (Fig. 6B). First, W200 in AaL (S in AiiB), a bulky residue, afects the position of the neighboring Y158 (Y151 in AiiB) and H120 (H113 in AiiB) (Fig. S6). As a result, the metal coordinating residue H120 as well as the metal cations adopt diferent positions in the two enzymes. In fact, metal coordination is afected and the distance between both metal cations varies greatly between the two enzymes (3.5 Å and 4.2 Å in AaL and AiiB, respectively, for the two phosphate-bound structures) (Fig. 6B). Concurrently, other metal coordinating residues adopt slightly diferent conformations such as H118 and D122, respectively H111 and D115 in AiiB (Fig. 6B). Other minor diferences can be observed. In the I237-containing loop (237-loop), residue I237 (in AaL) corresponds to V230 in AiiB (Fig. 6B). Some other residues lining the active site also difer. Indeed, as shown by the sequence alignment (Fig. 2), very few residues putatively involved in substrate binding are conserved. Noteworthy, most substitutions between AaL and AiiB (F19AiiB/W26AaL; A80AiiB/F87AaL; V230AiiB/I237AaL; N114AiiB/L121AaL; V15AiiB/M22AaL) suggest a more hydrophobic active site pocket for AaL. In particular, residues W26, F87 and I237 in AaL create a unique hydrophobic patch not present in other MLLs structures.

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 5 www.nature.com/scientificreports/

AaL bound to AaL bound to AaL bound to C6 Structure phosphate glycerol AHL DATA COLLECTION STATISTICS PDB ID 6CGY 6CH0 6CGZ Difraction source APS Argonne 23ID-B APS Argonne 23ID-D APS Argonne 23ID-D Wavelength (Å) 1.03323 1.03321 1.03321 Detector EIGER-16M PILATUS3–6M PILATUS3–6M Rotation range per image (°) 0.2 0.2 0.3 Total rotation range (°) 200 240 230 Space group C2 C2 C2 a = 111.72 a = 110.84, a = 145.94 b = 114.74 b = 113.50 b = 88.69 c = 79.97; c = 79.10; c = 97.68; Unit-cell parameters (Ǻ) α = 90.000 α = 90.000 α = 90.000 β = 109.778 β = 108.907 β = 128.240 γ = 90.000 γ = 90.000 γ = 90.000 Resolution range (Ǻ) (last bin) 1.65 (1.65–1.75) 2.15 (2.15–2.25) 1.8 (1.8–1.9) Total N° of refections (last bin) 431338 (70385) 217409 (28557) 378841 (57949) N° of unique refections (last bin) 111796 (17858) 50025 (6321) 90111 (13463) Completeness (%) (last bin) 98.2 (97.2) 99.3 (99.0) 99.2 (99.3) Redundancy 3.86 (3.94) 4.35 (4.52) 4.20 (4.30) 〈I/σ(I)〉 22.08 (3.16) 11.71 (2.73) 18.27 (2.71)

Rsym(%) 3.1 (51.1) 7.0 (61.5) 4.8 (52.2) CC(1/2) 100 (84.9) 99.8 (85.5) 99.9 (81.6) REFINEMENT STATISTICS Rfree/Rwork (%) 19.73/17.17 22.92/18.75 19.24/16.49 N° of total model atoms 7311 6979 7602 Ramachandran favored (%) 94.17 95.09 94.72 Generously allowed rotamers (%) 3.44 2.97 3.52 Ramachandran outliers (%) 2.38 1.94 1.76 Rmsd from ideal Bond lenghts (Ǻ) 0.030 0.017 0.023 Bond angles (°) 2.423 1.753 2.082

Table 3. Data collection and refnement statistics of AaL structures.

Comparison with AidC. Te entrance of the active site binuclear center in AidC is located on a diferent face of the enzyme, as compared with AiiA, AiiB, and AaL. AidC and AaL share a similar metal coordination (Fig. 6C), but due to the previously described reorganization of the AidC binding clef28, their binding clef are diferent. Noteworthy the conserved residue Y223 in MLLs, as well as in PLLs13, possibly involved in the catalytic mechanism, is replaced by an histidine residue in AidC35 (Fig. 6C).

Structure of AaL bound to a phosphate anion. Examination of one of the obtained AaL structure reveals the presence of a tetrahedral molecule in the active site, as clearly illustrated by the electronic density (Fig. 7A). Similarly to what was observed in the AiiB structure54, this density was attributed to a phosphate anion. In this structure, the α-metal is bridged with H118, H120, H198, D220 and two phosphate oxygen atoms, while the β-metal is bound to H123, H266, D122, D220 and two oxygen atoms from the phosphate anion. With the bound phosphate, the bridging water molecule is not present in the metal coordination. Te presence of this tetrahedral anion in the active site of lactonases might mimic the tetrahedral geometry of the lactone hydrolytic transition state2. Interestingly, the phosphate-bound AaL and AiiB structures difer in the way they coordinate the anion. In particular, in AiiB, N114 strongly interacts with one of the oxygen atom of the bound phosphate molecule (Fig. S7), demonstrating its ideal position to create polar interactions with molecules bound to the bi-metallic active site center, and thereby including lactones. However, in the case of AaL, the corresponding residue, a leucine, does not interact with the bound anion.

Structure of AaL bound to a glycerol molecule. AaL was also crystallized bound to a glycerol molecule. Glycerol may originate from the cryoprotectant solution that contains glycerol34. Similar cryoprotectant molecule (ethylene glycol) was previously observed to bind to bimetallic enzyme active sites in PTE58. Te molecule interact through two of its alcohol groups with the metal cations and with D122. In fact, an alcohol group is bridged with the α-metal, D122 and D220 at 2.0, 2.6 and 2.7 Å respectively and the second alcohol interacts with the β-metal (1.9 Å) (Fig. 7B). Surprisingly, whereas the metal coordination is very similar to other lactonases from the MLL family (AiiA, AiiB, AidC) and the PLL family (SsoPox, SisLac, VmoLac), and the good resolution of the

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 6 www.nature.com/scientificreports/

Figure 4. Structures of AaL and comparison to other known structures of MLLs. (A) AaL monomer shows a αβ/βα fold that contains two metal cations (pink spheres) and a bound phosphate anion (orange sticks). Te α-helices, β-sheets and loops are represented in red, yellow and green, respectively. (B) AaL dimer. Each monomer (in blue and gold) contains an active site with two cobalt cations (pink spheres) bound to a phosphate anion (orange sticks). (C) Comparison of AaL (grey) and AiiA (pink) monomers. (D) Superposition of AaL dimer (grey) with the AiiA (pink) monomer. (E) Superposition of the monomers AaL (grey) and AiiB (green). (F) Comparison of the dimers of AaL (grey) and AiiB (green). Te two enzymes show a similar dimerization mode. (G) Superposition of the monomer of AaL (grey) and AidC (cyan). (H) Superposition of AaL (grey) and AidC (cyan) dimers reveals their use of a diferent dimerization interface.

structure (2.15 Å), the putative metal cation-bridging water molecule is not visible in the electronic density maps. It is however difcult to conclude whether this molecule is present or not, since the observation of this water molecule can be difcult because of the strong density peaks corresponding to both metal cations, and the conserved coordination suggests that the conserved bridging water is present in AaL. When compared to the AaL structure bound to phosphate, both structures align very well, with the exception of the 237-loop (Fig. S8A). Additionally, the metal coordination is slightly diferent in the two structures. Indeed, the distance between the two metal cations is bigger in the glycerol bound structure, as compared to the phosphate bound one (4.0 Å and 3.5 Å, respectively). Tis lead to a shif in the position of the residues H120, H123, H266 and D220 which coordinates the metals (Fig. S8B).

Complex with C6-AHL. Afer soaking AaL crystals in the cryoprotectant solution supplemented with 20 mM of C6-AHL for 10 min, the structure of AaL bound to C6-AHL could be solved. C6-AHL is bound within

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 7 www.nature.com/scientificreports/

Figure 5. Active site accessibility in three diferent MLLs. (A) AiiA (PDB: 2A7M) active site cavity appears divided in two parts by two residues (F68 and E135), indicated by pink arrows. (B) AiiB (PDB: 2R2D) active site cavity. AiiB is bound to a phosphate molecule shown as orange sticks. (C) AaL (PDB: 6CGY) active site is bound to a phosphate molecule shown as orange sticks. Active site metal cations are shown as red spheres.

Figure 6. Comparison of AaL active site with AiiA, AiiB, and AidC. (A) Superposition of AaL (in grey sticks) and AiiA (in pink sticks) highlights a reorientation of Y223/Y194 and of the active site loop (237-loop in AaL). (B) Superposition of AaL (grey sticks) and AiiB (green sticks) reveals a diferent metal coordination geometry (AaL cations are shown as pink spheres, AiiB cations as grey spheres) and the substitution I237/V230. (C) Superposition of AaL (grey sticks) and AidC (cyan sticks) active sites. Active site residues occupies similar confgurations, with the exception of the Y223/H261 substitution, and residues on loops (P235/F274 and I237/ G276). Te phosphate-bound structure of AaL was used in these fgures. Metal cations (pink spheres for AaL, grey spheres for other enzymes) and the bridging water molecule are shown as well as the bound anions shown as sticks.

Figure 7. AaL active site in structures bound to phosphate and to glycerol. (A) Structure of AaL (green sticks) bound to phosphate (orange sticks) (PDB ID: 6CGY). 3 out of the 4 oxygen atoms of the phosphate molecule interact closely with the two metal cations α and β and as well as with D122 and Y223. (B) Structure of AaL (green sticks) bound to glycerol (pink sticks) (PDB ID: 6CHO). Te glycerol molecule interacts with the bimetallic active site via two oxygen atoms. Fourier diference maps (Fo-Fc) shown in green mesh are calculated by omitting the ligand from the model, and contoured at 5.0 σ (A) and 2.5 σ (B) level.

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 8 www.nature.com/scientificreports/

Figure 8. AaL complex with a C6-AHL. Active site residues (green sticks) are shown and interactions between the bound C6-AHL (purple sticks), metal cations, and the bridging water molecule are highlighted as black dashes and distances are indicated in Ångstroms. Te hydrophobic patch accommodating the AHL acyl chain is highlighted with a light blue background.

the active site (Fig. 8) with a modelled 0.7 occupancy (Fig. S9). As a result of soaking, the crystals belong to the same space group than the two other structures (C2), but the unit cell parameters are diferent (Table 3). Tis structure is overall very similar to the other structures of AaL, including in the metal coordination sphere. Metal cations are distant by 4.0 Å. A slight change of conformation of I237 is observable in the C6-AHL-bound structure as compared to the phosphate-bound structure (Fig. S10). Te lactone ring interacts with the bi-metallic active site (Fig. 8). Te carbonyl oxygen of the ring is bound to the α-cobalt (2.2 Å) and the esteric oxygen with the β-cobalt (2.1 Å). Te catalytic water is at 2.3 Å of the nucleophilic carbon that it will attack to degrade the substrate. Tis binding confguration is compatible with the previously observed binding mode of a hydrolytic product in AiiA54, but signifcant diferences are observable (Fig. S11). Notably, the bridging water molecule is interacting more closely with one of the metal cation (β-metal) than with the other metal (α-metal), as reported for PLLs21. Tis confguration is compatible with the proposed mechanism utilizing the activating bridging catalytic water as the attacking nucleophile to open up the lactone ring26. Te accommodation of the N-alkyl chain of the AHL is unique: whereas AiiA utilizes a wide and shallow crevice, where longer AHL residues can be stabilized by a phenylalanine clamp54, the binding clef in AaL is diferent. Te binding clef is much wider than in AiiA’s structure, very hydrophobic, revealing an additional hydrophobic cluster, or patch (Fig. 8). Te structure shows that the acyl chain interacts with the hydrophobic patch formed by W26, F87 and I237 that has no equivalent in other MLLs (Fig. S12), and this results in a diferent orientation of the alkyl chains in AaL, as compared to AiiA (Fig. S11). Te presence of unique hydrophobic patch may contrib- ute to the observed lower KM values of AaL. Additionally, the bound C6-AHL also interacts with M20, M22, M86, Y223, L121 and A157. Remarkably, as compared to the phosphate-bound structure, the 237-loop adopts a slight reorientation upon C6-AHL binding, including a reorientation of I237 side chain (Fig. S10). Te importance of the interaction of I237 with the acyl chain of C6-AHL is further evidenced by the decreased of its mobility in the C6-AHL-bound structure, as compared to the phosphate-bound structure, as evidenced by normalized thermal motion factors (B-factors) (Fig. S13). Conclusion Te quorum quenching lactonase AaL from the thermoacidophilic bacterium Alicyclobacillus acidoterrestris exhibits a very broad substrate range, being capable of hydrolyzing short and long chain AHLs with high profciencies. Tis broad substrate specifcity seems common to most of the MLL lactonases identifed thus far, including GcL30, MomL29, AidC28 or AiiA40. Te ability of AaL to hydrolyze AHLs results in its capacity to inhibit the bioflm formation of Acinetobacter baumannii. Additionally, AaL exhibits high catalytic profciency against δ-lactones and γ-lactones. Tis is noteworthy, because some γ-lactones are used as QS molecules in Streptomyces and Rhodococcus24,25. Furthermore, its promiscuous ability to degrade the phosphotriester paraoxon resonates with previous studies suggesting an evolutionary link between phosphotriesterases (PTEs) and lactonases13,46. Indeed, lactonases might be the progenitors of modern PTEs46, which are insecticide-degrading enzymes hypothesized to have emerged in the last 70 years to specifcally degrade man-made organophosphorus insecticides. Remarkably, for all three main lactonase families, namely the PLLs, the PONs, and the MLLs, closely related PTE representatives could be

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 9 www.nature.com/scientificreports/

identifed and characterized13,17,59–61. Te weak, promiscuous paraoxonase activity of AaL may confrm the previously observed catalytic relationships between PTEs and lactonases13,17,46. Lastly, the unusually low KM values of AaL correlates with the presence of a hydrophobic patch in the vicinity of the active site that is unique to AaL structure. Structural analysis of the structure bound to a C6-AHL molecule allows for the identifcation of the residues interacting with the acyl chain. In particular, a residue within this hydrophobic patch, I237, adopts slightly diferent conformations upon the binding of the C6-AHL molecule, suggesting a potential role in the AHL accommodation. Te use of lactonase with low KM values may be of particular interest to increase their quorum quenching abilities. Indeed, the majorities of quorum quenching enzymes identifed so far have high apparent dissociation constant values (100–1000 µM). Tese values contrast with the reported activation threshold of QS for numerous bacteria, in the range of ~5 nM62–64. Future investigations will reveal if the use of lactonases with lower KM values result in stronger quenching, and AaL will serve as an example to understand how lactonases can achieve low apparent dissociation constant values. Methods Sequence alignment. Te FASTA protein sequence of AaL was blasted against the non-redundant protein database to collect the sequences of previously characterized lactonases. The protein sequence of AaL (WP_021296945.1) contains 282 amino acids and has a molecular weight of 32.0 kDa. Te alignment of AaL and other AiiA-like lactonase representative was performed in MEGA65 by using the sofware MUSCLE38

Cloning, expression and purifcation of the protein AaL. Te gene encoding for AaL has been optimized for heterologous expression in Escherichia coli and synthesized by Genscript (Piscataway Township, New Jersey, USA). To performed production, AaL gene was built by the addition of a N-terminal afnity strep tag (WSHPQFEK) followed by a TEV cleavage site (ENLYFQS). AaL production was performed in E. coli BL21 (DE3)-pGro7/ GroEL strain (Takara). Te production was efected by using ZYP autoinducer media at 37 °C until cells reached the exponential growth phase (OD600nm). Ten, the cultures were transferred at 18 °C overnight and 0.2 mM CoCl2 were added, as well as 0.2% of L-arabinose to induce the chaperon. Te use of cobalt during induction and in subsequent purifcation steps was previously used for the purifcation of similar lactonases and was associated to increased induction and activity levels44. Te cells were collected by centrifugation at 4 °C (4400 g during 15 minutes) and resuspend in lysis bufer (50 mM HEPES pH 8.0, 150 mM NaCl, 0.2 mM CoCl2, 0.25 mg. ml−1 lysozyme, 0.1 mM phenylmethylsulfonyl (PMSF)). Afer 30 minutes in ice, cells were lysed by using a sonicator device (Q700 Sonicator Qsonica, USA) during 3 times 30 seconds (1 second pulse-on; 2 seconds pulse-of). Te lysate was fnally loaded on a Strep Trapp HP chromatography column (GE Healthcare) at room temperature. StrepTag encoding in AaL gene were thereafer cleaved by using TEV protease (Tobacco Etch Virus protease) at a ratio of 1/20 w/w, during 20 hours at 4 °C. Afer cleavage, protein sample were concentrated, fltered at 0.22 μm and loaded on a size exclusion column (Superdex 75 16/60, GE Healthcare). To identify and control the protein purity, samples were loaded on SDS-PAGE gel (Fig. S1).

Kinetic measurements. Reaction rates of AaL enzyme were monitored using a microplate reader (Synergy HTX, BioTek, USA) and the sofware Gen5.1 in 96-well plates of 5.8 mm path lengh for a 200 μl reaction at room temperature. Measurements were performed in at least triplicates. Te catalytic parameters were obtained by ftting the data to the Michaelis-Menten equation with the Graph-Pad Prism 5 sofware. In case were Vmax could not be reached, the catalytic efciency was obtained by ftting the linear part of Michaelis-Menten plot to a linear regression. Paraoxon assay. Te ethyl-paraoxon hydrolysis was performed in PTE bufer (50 mM HEPES pH 8.0, 150 mM NaCl, 0.2 mM CoCl2). Catalytic parameters were evaluated using 1.45 µM of enzyme and paraoxon concentrations ranging from 0 to 6 mM. Te assay were performed by measuring the production of paranitrophenolate −1 −1 23,42 anions (ε405nm = 17 000 M cm ) as previously described (Fig. 1). Lactonase assay. Te hydrolysis of lactones, via the opening of the lactone ring, increases the acidity of the solution by generating a proton. A pH indicator, the cresol purple (pKa 8.3 at 25 °C) was used to monitor the acidifcation of the medium induced by the hydrolysis of the lactone ring as previously described30,34. Te time course of the hydrolysis of the lactones was recorded at 577 nm in the activity bufer (2.5 mM Bicine pH 8.3, 150 mM NaCl, 0.2 mM cresol purple, 0.5% DMSO and with 0.2 mM CoCl2). Te assays were performed using AHLs as substrates: C4-AHL, C6-AHL, C10-AHL, 3-oxo-C12-AHL and using other lactones: γ-butyrolactone, γ-heptanolide, δ-valerolactone, δ-decalactone as substrates (Table S1). Te concentration range for the diferent lactones varied between 0.001 mM and 1.5 mM. Te enzyme concentrations used for kinetic experiments were adjusted as a function of observed velocities to optimize the quality of the acquired data sets. For each tested substrate, a specifc concentration of enzyme, between 1.45 µM and 0.73 µM, was used. Te kinetics parameters of AaL against AHLs C4-AHL, C6-AHL, C10-AHL, 3-oxo-C12-AHL were previously published34.

Data collection, structure resolution and refnement. Crystals of AaL bound to phosphate and bound to glycerol were obtained in conditions previously reported34. Crystals bound to C6-AHL were obtained afer soaking the crystals in a cryoprotectant solution supplemented with 20 mM C6-AHL for 10 min. AaL structure was collected and obtained by X-ray difraction intensities at APS-Argonne on the beamline 23ID-B (Lemont, Illinois, United states) using an EIGER detector. Te collection were performed at 1.03323 Ǻ wavelength and 1000 images were collected (0.2 s exposure; 0.2° step). Te XDS sofware package66 was used to integrate and scale X-ray difraction data. Te closest known structure to AaL, AiiB from Agrobacterium tumefaciens (43% sequence identity)54 (PDB: 2R2D) was used as a template for molecular replacement procedure using MOLREP67. Buccaneer sofware68 was used to automatically build a partially model of AaL that was subsequently improved manually

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 10 www.nature.com/scientificreports/

using the sofware Coot69. Refnement cycles were performed using REFMAC70 to improve the structural model and the fnal refnement statistics are available in Table 3.

Anomalous X-ray scattering data. In order to determine the chemical nature of the two active-site metals, an anomalous data set was collected. Because the bufers used during the purifcation step contained salts, and because cobalt cations are known to be present in active sites of lactonases21, we collected a 2.4 Å resolution set at higher energy (7.715 keV) than the Co-K absorption edge (7.7089 KeV). (Table S2 and Fig. S2)

Structure analysis. Te crystals structures of AaL (PDB 6CGY), AiiA (PDB 2A7M and 3DHB), AiiB (PDB 2R2D) and AidC (PDB 4ZO2) were compared. PyMOL sofware was used to analyze, compare and illustrate our data71. Te dimer interface surface together with the number of hydrogen bonds and salt bridges were computed using PISA32 server and default parameters72,73. Te determination of the root mean square deviations (r.m.s.d.) were calculated on α -carbon using the align command under the Swiss-PdbViewer74.

Bioflm inhibition assay. Te bacterial strain Acinetobacter baumannii ATCC 19606TM was used in this study. Te bacteria were grown in nutrient broth (NB) medium (Fisher Scientifc) as® instructed by ATCC. Te bioflm assays were carried out in MOPS minimal medium made according to Neidhardt et al. protocol75 (50 mM MOPS, 4 mM Tricine, 50 mM NaCl, 1 mM K2HSO4, 50 mM MgCl2, 10 mM CaCl2, 0.3 mM (NH4)6Mo7O24, 40 mM H3BO3, 3 mM Co(OAc)2, 1 mM CuSO4, 8 mM MnSO4, 1 mM ZnSO4, 25 mM sodium glutamate 15 mM NH4Cl, 4 mM K2HPO4, 5 µM FeSO4 [pH 7.0]). A single colony of A. baumannii was inoculated into 5 ml of NB medium and grown at 37 °C, 200 rpm shaking for 6 hr as pre-culture. Te pre-culture was diluted 1:1000 in MOPS minimal medium and inoculated into a sterile polystyrene 96-well round-bottom plate (Costar) for bioflm assay with diferent concentration of the lactonase. Te enzyme concentrations were prepared by serial dilutions and added to the 96-well plate in triplicates. Te plate was incubated at 37 °C with continuous agitation and bioflm was assayed 16 hr post inoculation by crystal violet staining. Te crystal violet staining was carried out as reported by O’Toole (2011)76. Briefy, the unattached cells were removed by gently pipetting out into another plate and the cell density was measured by absorbance at 600 nm using a plate reader (Synergy HTX multi-mode reader, BioTek). Te culture plate was then washed by gently immersing it in tub full of distilled water and gently shaking out the water and blotting the plate dry. Te bioflm was allowed to dry at room temperature for half an hour and then 200 µl of 0.1% crystal violet solution was added to each well. Te crystal violet was discarded and the wells were washed as mentioned earlier. Te crystal violet was solubilized using 30% acetone. 200 µl from each well were transferred to a fresh fat-bottom 96-well plate (Fisherbrand, Fisher Scientifc) and the absorbance was measured at 550 nm. References 1. Parsek, M. R., Val, D. L., Hanzelka, B. L., Cronan, J. E. & Greenberg, E. P. Acyl homoserine-lactone quorum-sensing signal generation. Proc. Natl. Acad. Sci. 96, 4360–4365 (1999). 2. Hawver, L. A., Jung, S. A. & Ng, W.-L. Specifcity and complexity in bacterial quorum-sensing systems. FEMS Microbiol. Rev. 40, 738–752 (2016). 3. Dong, Y.-H. & Zhang, L.-H. Quorum sensing and quorum-quenching enzymes. J. Microbiol. Seoul Korea 43(Spec No), 101–109 (2005). 4. Fetzner, S. Quorum quenching enzymes. J. Biotechnol. 201, 2–14 (2015). 5. Bzdrenga, J. et al. Biotechnological applications of quorum quenching enzymes. Chem. Biol. Interact. 267, 104–115 (2017). 6. Zhang, Y., Brackman, G. & Coenye, T. Pitfalls associated with evaluating enzymatic quorum quenching activity: the case of MomL and its efect on Pseudomonas aeruginosa and Acinetobacter baumannii bioflms. PeerJ 5, e3251 (2017). 7. Augustine, N., Kumar, P. & Tomas, S. Inhibition of Vibrio cholerae bioflm by AiiA enzyme produced from Bacillus spp. Arch. Microbiol. 192, 1019–1022 (2010). 8. Vinoj, G., Vaseeharan, B., Tomas, S., Spiers, A. J. & Shanthi, S. Quorum-quenching activity of the AHL-lactonase from Bacillus licheniformis DAHB1 inhibits Vibrio bioflm formation in vitro and reduces shrimp intestinal colonisation and mortality. Mar. Biotechnol. N. Y. N 16, 707–715 (2014). 9. Tang, K. & Zhang, X.-H. Quorum Quenching Agents: Resources for Antivirulence Terapy. Mar. Drugs 12, 3245–3282 (2014). 10. Cao, Y. et al. Orally Administered Termostable N-Acyl Homoserine Lactonase from Bacillus sp. Strain AI96 Attenuates Aeromonas hydrophila Infection in Zebrafsh. Appl. Environ. Microbiol. 78, 1899–1908 (2012). 11. Dobretsov, S., Teplitski, M. & Paul, V. Mini-review: quorum sensing in the marine environment and its relationship to biofouling. Biofouling 25, 413–427 (2009). 12. LaSarre, B. & Federle, M. J. Exploiting quorum sensing to confuse bacterial pathogens. Microbiol. Mol. Biol. Rev. 77, 73–111 (2013). 13. Elias, M. & Tawfk, D. S. Divergence and Convergence in Enzyme Evolution: Parallel Evolution of Paraoxonases from Quorum- quenching Lactonases. J Biol Chem 287, 11–20 (2012). 14. Amara, N., Krom, B. P., Kaufmann, G. F. & Meijler, M. M. Macromolecular inhibition of quorum sensing: enzymes, antibodies, and beyond. Chem. Rev. 111, 195–208 (2011). 15. Yang, F. et al. Quorum quenching enzyme activity is widely conserved in the sera of mammalian species. FEBS Lett 579, 3713–7 (2005). 16. Khersonsky, O. & Tawfk, D. S. Structure-reactivity studies of serum paraoxonase PON1 suggest that its native activity is lactonase. Biochemistry (Mosc.) 44, 6371–82 (2005). 17. Ben-David, M. et al. Catalytic versatility and backups in enzyme active sites: the case of serum paraoxonase 1. J Mol Biol 418, 181–96 (2012). 18. Ben-David, M. et al. Catalytic metal ion rearrangements underline promiscuity and evolvability of a metalloenzyme. J Mol Biol 425, 1028–38 (2013). 19. Harel, M. et al. Structure and evolution of the serum paraoxonase family of detoxifying and anti-atherosclerotic enzymes. Nat Struct Mol Biol 11, 412–9 (2004). 20. Bar-Rogovsky, H., Hugenmatter, A. & Tawfk, D. S. Te Evolutionary Origins of Detoxifying Enzymes Te MAMMALIAN SERUM PARAOXONASES (PONs) Relate To Bacterial Homoserine Lactonases. J. Biol. Chem. 288, 23914–23927 (2013). 21. Elias, M. et al. Structural basis for natural lactonase and promiscuous phosphotriesterase activities. J Mol Biol 379, 1017–28 (2008).

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 11 www.nature.com/scientificreports/

22. Hiblot, J., Bzdrenga, J., Champion, C., Chabriere, E. & Elias, M. Crystal structure of VmoLac, a tentative quorum quenching lactonase from the extremophilic crenarchaeon Vulcanisaeta moutnovskia. Sci Rep 5, 8372 (2015). 23. Hiblot, J., Gotthard, G., Chabriere, E. & Elias, M. Structural and Enzymatic characterization of the lactonase SisLac from Sulfolobus islandicus. PLoS ONE 7, (2012). 24. Safari, M., Amache, R., Esmaeilishirazifard, E. & Keshavarz, T. Microbial metabolism of quorum-sensing molecules acylhomoserine lactones, γ-heptalactone and other lactones. Appl. Microbiol. Biotechnol. 98, 3401–3412 (2014). 25. Ceniceros, A., Dijkhuizen, L. & Petrusma, M. Molecular characterization of a Rhodococcus jostii RHA1 γ-butyrolactone (-like) signalling molecule and its main biosynthesis gene gblA. Sci. Rep. 7, 17743 (2017). 26. Momb, J. et al. Mechanism of the Quorum-Quenching Lactonase (AiiA) from Bacillus thuringiensis. 2. Substrate Modeling and Active Site Mutations†. Biochemistry (Mosc.) 47, 7715–7725 (2008). 27. Liu, D. et al. Structure and specifcity of a quorum-quenching lactonase (AiiB) from Agrobacterium tumefaciens. Biochemistry (Mosc.) 46, 11789–99 (2007). 28. Mascarenhas, R. et al. Structural and biochemical characterization of AidC, a quorum-quenching lactonase with atypical selectivity. Biochemistry (Mosc.) 54, 4342–4353 (2015). 29. Tang, K. et al. MomL, a novel marine-derived N-acyl homoserine lactonase from Muricauda olearia. Appl. Environ. Microbiol. 81, 774–782 (2015). 30. Bergonzi, C., Schwab, M. & Elias, M. The quorum-quenching lactonase from Geobacillus caldoxylosilyticus: purification, characterization, crystallization and crystallographic analysis. Acta Crystallogr F Struct Biol Commun 72, 681–6 (2016). 31. Liu, D. et al. Three-dimensional structure of the quorum-quenching N-acyl homoserine lactone hydrolase from Bacillus thuringiensis. Proc. Natl. Acad. Sci. USA 102, 11882–11887 (2005). 32. Liu, D. et al. Mechanism of the quorum-quenching lactonase (AiiA) from Bacillus thuringiensis. 1. Product-bound structures. Biochemistry (Mosc.) 47, 7706–14 (2008). 33. Dong, Y.-H. et al. Quenching quorum-sensing-dependent bacterial infection by an N-acyl homoserine lactonase. Nature 411, 813–817 (2001). 34. Bergonzi, C., Schwab, M., Chabriere, E. & Elias, M. The quorum-quenching lactonase from Alicyclobacter acidoterrestris: purifcation, kinetic characterization, crystallization and crystallographic analysis. Acta Crystallogr. Sect. F 73, 476–480 (2017). 35. Wang, W.-Z., Morohoshi, T., Someya, N. & Ikeda, T. AidC, a Novel N-Acylhomoserine Lactonase from the Potato Root-Associated Cytophaga-Flavobacteria-Bacteroides (CFB) Group Bacterium Chryseobacterium sp. Strain StRB126. Appl. Environ. Microbiol. 78, 7985–7992 (2012). 36. Hiblot, J., Gotthard, G., Elias, M. & Chabriere, E. Diferential active site loop conformations mediate promiscuous activities in the lactonase SsoPox. PLoS One 8, e75272 (2013). 37. Chow, J. Y. et al. Directed evolution of a thermostable quorum-quenching lactonase from the amidohydrolase superfamily. J Biol Chem 285, 40911–20 (2010). 38. Edgar, R. C. MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics 5, 113 (2004). 39. Kim, M. H. et al. Te molecular structure and catalytic mechanism of a quorum-quenching N-acyl-L-homoserine lactone hydrolase. Proc. Natl. Acad. Sci. USA 102, 17606–17611 (2005). 40. Wang, L.-H., Weng, L.-X., Dong, Y.-H. & Zhang, L.-H. Specificity and enzyme kinetics of the quorum-quenching N-Acyl homoserine lactone lactonase (AHL-lactonase). J. Biol. Chem. 279, 13645–13651 (2004). 41. Bzdrenga, J. et al. SacPox from the thermoacidophilic crenarchaeon Sulfolobus acidocaldarius is a profcient lactonase. BMC Res Notes 7, 333 (2014). 42. Hiblot, J., Gotthard, G., Chabriere, E. & Elias, M. Characterisation of the organophosphate hydrolase catalytic activity of SsoPox. Sci. Rep. 2, (2012). 43. Du, Y.-L., Shen, X.-L., Yu, P., Bai, L.-Q. & Li, Y.-Q. Gamma-Butyrolactone Regulatory System of Streptomyces chattanoogensis Links Nutrient Utilization, Metabolism, and Development. Appl. Environ. Microbiol. 77, 8415–8426 (2011). 44. Merone, L., Mandrich, L., Rossi, M. & Manco, G. A thermostable phosphotriesterase from the archaeon Sulfolobus solfataricus: cloning, overexpression and properties. Extrem. Life Extreme Cond. 9, 297–305 (2005). 45. Elias, M. et al. Crystallization and preliminary X-ray diffraction analysis of the hyperthermophilic Sulfolobus solfataricus phosphotriesterase. Acta Crystallogr Sect F Struct Biol Cryst Commun 63, 553–5 (2007). 46. Afriat-Jurnou, L., Jackson, C. J. & Tawfik, D. S. Reconstructing a missing link in the evolution of a recently diverged phosphotriesterase by active-site loop remodeling. Biochemistry (Mosc.) 51, 6047–55 (2012). 47. Anbazhagan, D. et al. Detection of quorum sensing signal molecules and identifcation of an autoinducer synthase gene among bioflm forming clinical isolates of Acinetobacter spp. PloS One 7, e36696 (2012). 48. López, M. et al. Quorum sensing network in clinical strains of A. baumannii: AidA is a new quorum quenching enzyme. PloS One 12, e0174454 (2017). 49. Chow, J. Y., Yang, Y., Tay, S. B., Chua, K. L. & Yew, W. S. Disruption of bioflm formation by the human pathogen Acinetobacter baumannii using engineered quorum-quenching lactonases. Antimicrob. Agents Chemother. 58, 1802–1805 (2014). 50. Jacquet, P. et al. Rational engineering of a native hyperthermostable lactonase into a broad spectrum phosphotriesterase. Sci. Rep. 7, 16745 (2017). 51. Guendouze, A. et al. Efect of Quorum Quenching Lactonase in Clinical Isolates of Pseudomonas aeruginosa and Comparison with Quorum Sensing Inhibitors. Front. Microbiol. 8, (2017). 52. Hraiech, S. et al. Inhaled Lactonase Reduces Pseudomonas aeruginosa Quorum Sensing and Mortality in Rat Pneumonia. PLoS One 9, e107125 (2014). 53. García-Contreras, R. et al. Resistance to the quorum-quenching compounds brominated furanone C-30 and 5-fuorouracil in Pseudomonas aeruginosa clinical isolates. Pathog. Dis. 68, 8–11 (2013). 54. Liu, C. F. et al. A phenylalanine clamp controls substrate specifcity in the quorum-quenching metallo-γ-lactonase from Bacillus thuringiensis. Biochemistry (Mosc.) 52, 1603–1610 (2013). 55. Vieille, C. & Zeikus, G. J. Hyperthermophilic enzymes: sources, uses, and molecular mechanisms for thermostability. Microbiol. Mol. Biol. Rev. MMBR 65, 1–43 (2001). 56. Omburo, G. A., Mullins, L. S. & Raushel, F. M. Structural characterization of the divalent cation sites of bacterial phosphotriesterase by 113Cd NMR spectroscopy. Biochemistry (Mosc.) 32, 9148–55 (1993). 57. Charendof, M. N., Shah, H. P. & Briggs, J. M. New insights into the binding and catalytic mechanisms of Bacillus thuringiensis lactonase: insights into B. thuringiensis AiiA mechanism. PloS One 8, e75395 (2013). 58. Jackson, C., Kim, H. K., Carr, P. D., Liu, J. W. & Ollis, D. L. Te structure of an enzyme-product complex reveals the critical role of a terminal hydroxide nucleophile in the bacterial phosphotriesterase mechanism. Biochim Biophys Acta 1752, 56–64 (2005). 59. Gotthard, G., Hiblot, J., Gonzalez, D., Elias, M. & Chabriere, E. Structural and enzymatic characterization of the phosphotriesterase OPHC2 from Pseudomonas pseudoalcaligenes. PLoS One 8, e77995 (2013). 60. Gotthard, G., Hiblot, J., Gonzalez, D., Chabriere, E. & Elias, M. Crystallization and preliminary X-ray difraction analysis of the organophosphorus hydrolase OPHC2 from Pseudomonas pseudoalcaligenes. Acta Crystallogr Sect F Struct Biol Cryst Commun 69, 73–6 (2013). 61. Dong, Y. J. et al. Crystal structure of methyl parathion hydrolase from Pseudomonas sp. WBC-3. J Mol Biol 353, 655–63 (2005).

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 12 www.nature.com/scientificreports/

62. Holden, M. T. et al. Quorum‐sensing cross talk: isolation and chemical characterization of cyclic dipeptides from Pseudomonas aeruginosa and other gram‐negative bacteria. Mol. Microbiol. 33, 1254–1266 (1999). 63. Trovato, A. et al. Quorum vs. difusion sensing: a quantitative analysis of the relevance of absorbing or refecting boundaries. FEMS Microbiol. Lett. 352, 198–203 (2014). 64. Michael, B., Smith, J. N., Swif, S., Hefron, F. & Ahmer, B. M. SdiA of Salmonella enterica is a LuxR homolog that detects mixed microbial communities. J. Bacteriol. 183, 5733–5742 (2001). 65. Tamura, K. et al. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol. Biol. Evol. 28, 2731–2739 (2011). 66. Kabsch, W. Automatic processing of rotation difraction data from crystals of initially unknown symmetry land cell constants. J Appl Cryst 26, 795–800 (1993). 67. Vagin, A. & Teplyakov, A. Molecular replacement with MOLREP. Acta Crystallogr. D Biol. Crystallogr. 66, 22–25 (2010). 68. Cowtan, K. Te Buccaneer sofware for automated model building. 1. Tracing protein chains. Acta Crystallogr. D Biol. Crystallogr. 62, 1002–1011 (2006). 69. Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr Biol Crystallogr 60, 2126–32 (2004). 70. Murshudov, G. N. et al. REFMAC5 for the refnement of macromolecular crystal structures. Acta Crystallogr. D Biol. Crystallogr. 67, 355–367 (2011). 71. DeLano, W. L. Te PyMOL Molecular Graphics System. DeLano Sci. San Carlos CA USA (2002). 72. Krissinel, E. & Henrick, K. Detection of Protein Assemblies in Crystals. inComputational Life Sciences 163–174 (Springer, Berlin, Heidelberg). https://doi.org/10.1007/11560500_15 2005 73. Krissinel, E. & Henrick, K. Inference of macromolecular assemblies from crystalline state. J. Mol. Biol. 372, 774–797 (2007). 74. Kaplan, W. & Littlejohn, T. G. Swiss-PDB Viewer (Deep View). Brief. Bioinform. 2, 195–197 (2001). 75. Neidhardt, F. C., Bloch, P. L. & Smith, D. F. Culture medium for enterobacteria. J. Bacteriol. 119, 736–747 (1974). 76. O’Toole, G. A. Microtiter Dish Bioflm Formation Assay. J. Vis. Exp. JoVE https://doi.org/10.3791/2437 (2011). 77. Bzdrenga, J. et al. SacPox from the thermoacidophilic crenarchaeon Sulfolobus acidocaldarius is a profcient lactonase. BMC Res. Notes 7, 333 (2014). 78. Clevenger, K. D., Wu, R., Er, J. A. V., Liu, D. & Fast, W. Rational Design of a Transition State Analogue with Picomolar Afnity for Pseudomonas aeruginosa PvdQ, a Siderophore Biosynthetic Enzyme. ACS Chem. Biol. 8, 2192–2200 (2013). Acknowledgements Tis work was supported by the MnDrive Initiative, the MnDrive demonstration grant, the BTI Biocatalysis Initiative, and the BARD grant IS-4960–16 FR to M.E. It was also supported by the French Government under the «Investissements d’avenir» (Investments for the Future) program managed by the Agence Nationale de la Recherche (ANR, fr: National Agency for Research; 10-IAHU-03) to E.C. and D.D. We thank the Advanced Photon Source and the beamline staf for access and support. Author Contributions C.B, M.S., T.N., M.E., performed the experiments, C.B, M.S., T.N., D.D., E.C., M.E. analyzed the data, C.B., M.E. wrote the manuscript and prepared the fgures, M.E. designed the experiments. Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018-28988-5. Competing Interests: Te authors declare no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. Te images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

ScIenTIfIc REPOrTS | (2018)8:11262 | DOI:10.1038/s41598-018-28988-5 13